Abstract

OBJECTIVE

A 6-year study was performed to determine the circumstances, causes, and outcomes of concussions in the National Football League.

METHODS

Between 1996 and 2001, the epidemiological features of concussions were recorded by National Football League teams with a standardized reporting form. Symptoms were reported and grouped as general symptoms, cranial nerve symptoms, memory or cognitive problems, somatic complaints, and loss of consciousness. The medical actions taken were recorded. In total, 787 game-related cases were reported, with information on the players involved, type of helmet impact, symptoms, medical actions, and days lost. Concussion risks were calculated according to player game positions.

RESULTS

There were 0.41 concussions per National Football League game. The relative risk was highest for quarterbacks (1.62 concussions/100 game-positions), followed by wide receivers (1.23 concussions/100 game-positions), tight ends (0.94 concussion/100 game-positions), and defensive secondaries (0.93 concussion/100 game-positions). The majority of concussions (67.7%) involved impact by another player's helmet. The remainder involved impact by other body regions of the striking player (20.9%) or ground contact (11.4%). The three most common symptoms of mild traumatic brain injury were headaches (55.0%), dizziness (41.8%), and blurred vision (16.3%). The most common signs noted in physical examinations were problems with immediate recall (25.5%), retrograde amnesia (18.0%), and information-processing problems (17.5%). In 58 of the reported cases (9.3%), the players lost consciousness; 19 players (2.4%) were hospitalized. A total of 92% of concussed players returned to practice in less than 7 days, but that value decreased to 69% with unconsciousness.

CONCLUSION

The professional football players most vulnerable to concussions are quarterbacks, wide receivers, and defensive secondaries. Concussions involved 2.74 symptoms/injury, and players were generally removed from the game. More than one-half of the players returned to play within 1 day, and symptoms resolved in a short time in the vast majority of cases.

Mild traumatic brain injury (MTBI) or concussion has been defined as a traumatically induced alteration in neural function, which may or may not involve loss of consciousness (8,39). For the purposes of this study, a broad but specific definition of concussion was adopted, as stated below.

MTBI is a major public health problem in the United States, with an estimated annual incidence of 160 to 375 cases/100,000 persons (23). The Centers for Disease Control and Prevention (4) estimated that the number of MTBIs has reached 300,000 cases/yr in all sports. Mild head injury has received a great deal of publicity in recent years because of its prevalence in many sports with contact or the potential for collisions; it has become a major issue in athletics. Powell and Barber-Foss (39) estimated that 3.9 to 7.7% of high school and college athletes sustain MTBIs each year, in all sports.

At the professional level, publicity has increased attention regarding MTBIs, because of repeated concussions among a number of high-profile professional football players. Until recently, however, little was known regarding the long-term sequelae of MTBIs. Although research has been conducted and progress has been made, some research has involved anecdotally based information and there have been few scientific data available regarding the symptoms of MTBI (concussion) among professional football players. There is also a paucity of epidemiological information available regarding the natural history of concussion in this population.

This article is the result of a multiyear effort by the National Football League (NFL) to address the deficiencies in our knowledge regarding concussion. The specific purpose is to analyze patterns of concussions and trends with a special epidemiological database, in the hope of identifying the most vulnerable positions and most frequent symptoms. The research efforts are ultimately aimed at improving the safety of football for players at all levels (professional, college, high school, and pee-wee), the safety of other contact sports for all participants, and the general public safety during nonathletic endeavors that place individuals at risk of head injury.

In 1994, the NFL formed the MTBI Committee in response to safety concerns regarding head injuries. Background information on the committee was reported by Pellman (36). The committee consisted of medical and scientific experts in the areas of traumatic brain injury, basic science research, and epidemiology. The mission of the committee was to scientifically investigate the subject of MTBIs in the NFL. The committee examined the existing literature and spent time interviewing a variety of experts to identify what was known and accepted regarding the natural history of MTBIs and to determine what areas required additional research. Committee members were on-field physicians and athletic trainers with experience recognizing and treating MTBIs.

After an initial review, the committee identified two areas that required attention. First, there was a strong need to be able to monitor the frequency of MTBIs in the NFL and to more specifically identify the clinical symptoms associated with concussions. Second, the MTBI Committee undertook a series of research projects aimed at defining concussion biomechanics in professional football. On the basis of analyses of game videos of MTBIs and laboratory reconstructions of the impacts with instrumented test dummies, the biomechanical features of concussions were determined for professional players. Those findings included the acceleration severity of the injuries (38) and the location and direction of the impacts (37). This article discusses the epidemiological features of MTBI in professional football and combines those data with the clinical symptoms present at the time of injury.

The NFL has collected epidemiological data on player injuries since 1980. The data included “concussions,” regardless of time lost from participation. The committee analyzed those data and identified a need to collect clinically oriented information regarding head injuries among the league's players. In addition to the data being recorded by the NFL Injury Surveillance System, the specific clinical symptoms evident at the time of injury needed to be documented.

Information regarding the circumstances at the time of injury and the clinical nature of the injury was also essential for determining the natural history of MTBIs. The committee agreed that data on the diagnosis and management of MTBIs should be recorded by the team physicians, with a standardized form. The program began with the 1995 NFL season and has continued in the subsequent years.

At the beginning of the project in 1995, the committee recognized a need to develop an all-inclusive definition of MTBI, so that all of the team physicians and athletic trainers could easily recognize the symptoms of reportable injuries. It was understood that any player with a recognized symptom of head injury, no matter how minor, should be included in the study. In the initial year of the project, a reportable MTBI was characterized by an “altered mental state regardless of duration and/or altered memory, regardless of duration or content, that resulted from trauma and occurred in an NFL practice or game.” Feedback from physicians and the committee's analysis of the project data during the first year of the project resulted in the development and modification of report forms and enhancement of the definition of reportable MTBIs.

The definition introduced by the committee in 1996 and used for the remainder of the study was as follows. A reportable MTBI was defined as a traumatically induced alteration in brain function, manifested by 1) alteration of awareness or consciousness, including but not limited to being dinged, dazed, stunned, woozy, foggy, or amnesic or, less commonly, being rendered unconscious or experiencing seizures, and 2) signs and symptoms commonly associated with postconcussion syndrome, including persistent headaches, vertigo, lightheadedness, loss of balance, unsteadiness, syncope, near-syncope, cognitive dysfunction, memory disturbances, hearing loss, tinnitus, blurred vision, diplopia, visual loss, personality changes, drowsiness, lethargy, fatigue, and inability to perform usual daily activities. The definition of concussion used by the MTBI Committee was a natural extension of a much earlier definition proposed by the Ad Hoc Committee to Study Head Injury Nomenclature of the Congress of Neurological Surgeons, which in 1966 defined concussion as “a clinical syndrome characterized by immediate transient impairment of neural function such as alteration of consciousness, disturbance of vision, equilibrium, etc., due to mechanical forces”(8a).

PATIENTS AND METHODS

Data Collection

Since 1980, the NFL Injury Surveillance System has provided a mechanism for athletic trainers and team physicians to record data on injured players and circumstances surrounding injuries. The system requires that each team record data on all concussions that occur, regardless of the amount of time lost to participation because of the injury. The recorded data include the player time lost, the player position, the activity of the player at the time of the injury, and the nature of the activities of the team. For the current research project, the data set was modified to include the equipment being worn, the mechanism of injury (e.g., head-to-head impact), the facemask, the chin strap, the type of mouthpiece (if worn), and the approximate location of the impact on the helmet.

In addition to data recorded by the athletic trainers, the committee devised a simple form regarding observed and reported symptoms for the individual team physicians to complete when they evaluated, in initial and follow-up visits, players who had sustained MTBIs. Players’ names were not included on the forms, to maintain confidentiality. The players were identified with the last six digits of their Social Security numbers; with this coding system, it was possible to merge the NFL Injury Surveillance System data with data on the clinical evaluations of injuries.

Symptoms

MTBI is a clinical syndrome that may present with a wide range of symptoms, many of which are rather nonspecific and can be associated with other clinical diagnoses. The categories in Table 1 represent groupings of 34 of the most common symptoms of concussion, including unconsciousness. The list was developed by MTBI Committee members who are NFL team physicians, as well as MTBI Committee consultants with special expertise in the fields of sport neuropsychology and sport neurology. Symptoms of concussion were grouped into six categories, as follows: 1) general symptoms, 2) cranial nerve symptoms, 3) memory problems, 4) cognitive problems, 5) somatic complaints, and 6) unconsciousness. These symptoms are consistent with findings previously noted after traumatic brain injury (3,14,20,21,28,33,41,45).

TABLE 1.

Signs and symptoms recorded on the mild traumatic brain injury forma

Players spontaneously reported many of the symptoms, but the complete symptom complex, including mental status (retrograde amnesia, anterograde amnesia, and problems with information processing, attention, and immediate recall), was assessed with physician questions. The committee did not distribute uniform testing instruments to the team physicians and instead left the assessment of these symptoms to the discretion of individual team physicians.

Efforts to Improve Compliance

The NFL, through the commissioner, strongly encouraged all team physicians to complete and return the project's forms whenever they examined players with head injuries. The project was designed to record information regarding the injuries. The initial form contained questions regarding the initial symptoms of the MTBI, the physical examination findings, the initial treatment, the tests ordered, and the disposition regarding the return to play. Physicians completed follow-up forms that documented the symptoms noted, the tests ordered and their results, and the time before return to full participation. Physicians used their own evaluation procedures to treat the injuries. The committee had neither the authority nor the inclination to impose outside medical decision-making on the medical staffs of the individual teams. The individual team physicians were to complete the initial and follow-up forms on the basis of their clinical findings.

To improve compliance, the forms were designed for ease of completion and the data were limited to the points that would provide the strongest and most consistent information on MTBI. After being completed, the data forms were sent to the NFL epidemiologist and entered into a database with a blinded coding system, to maintain the anonymity of the players. When an initial evaluation form was submitted but the follow-up forms were not, committee members contacted the team athletic trainers and doctors directly, to remind them to submit the follow-up forms.

In biannual meetings, the committee monitored the data and discussed the findings. Approximately 1.5 to 2 years after the initiation of the project, the forms were modified according to the findings of the analysis. For example, Glasgow Coma Scale scores were removed after the first year and data on loss of consciousness were added. Before the 1999 season, the forms were modified again, to remove fields that were determined to be of little value in the analysis. For example, the physical examination fields were modified so that the physicians could better describe their findings. To enhance other aspects of the committee's work, the identification and dates of neuropsychological tests were included on the physician's form.

Educational symposia based on the medical literature and the clinical backgrounds of committee members were held throughout the NFL, to increase awareness of the implications of even the mildest MTBIs and to promote compliance, to ensure the longevity of players’ careers. With promulgation of the definition throughout the NFL, the committee considered it more likely that all players with MTBIs would be included in the data collection efforts. The committee held numerous educational symposia on MTBI for medical personnel for all NFL teams, thus increasing awareness of MTBI and compliance with the study of MTBI.

Final MTBI Epidemiological Form

After a review of the data from the 1995 season, the recording forms were modified to improve clarity and the definition of a reportable event was extended. Because of these significant changes in the program, the 1995 data were considered pilot data and were not included in the analysis. With the specific limitations of the data, the project has complete injury data, initial clinical evaluation data, and follow-up evaluation data for 787 MTBIs that occurred in preseason, regular, or playoff games during the six professional football seasons from 1996 to 2001. There were 100 additional MTBI cases from practice sessions, which are included in the database but are not analyzed here because the risks are being considered according to player position and play type in games. Furthermore, unconsciousness data were not initially collected and, when such data were added to the standardized form, there were cases in which no determination was made by the presiding physician. Determinations were made in 623 cases.

Quality Assurances

The clinical evaluation forms were designed to include an initial evaluation form, to be completed at the time of injury, and a follow-up evaluation form, to be completed each time the physician evaluated the player until his return to full participation. The data forms were sent to the NFL epidemiologist and entered into the database. The MTBI evaluation forms were logged in and scanned into a database file with a commercial software program (Teleforms, Cardiff, CA). During the data logging, the individual forms were manually reviewed. Each form was scanned into a temporary database and verified before it was entered into the final database. Fields that were incomplete or inconsistent triggered a follow-up contact with the team athletic trainer or physician for data verification. When the MTBI evaluation data were merged with the injury surveillance data, the data were again reviewed and verified. The final database includes information from the initial and follow-up evaluation forms submitted by the team physicians.

Data Analyses

This analysis involved data on MTBIs that occurred during games in the 1996 to 2001 NFL seasons. The data reflect only injuries that occurred during preseason, regular, or playoff games. The elimination of injuries that occurred during practice sessions allowed the analysis to focus on NFL games and the activities of the players at the time of injury. It also allowed a calculation of risks on the basis of player positions.

Each MTBI was analyzed as an independent event, as a first step toward understanding concussions. Summary measures are reported with 95% confidence intervals. Nonoverlapping intervals may be interpreted as significantly different at α = 0.05. Exploratory analyses were performed to determine factors associated with concussion outcomes (i.e., number of days out of play and hospitalization). Factors associated with the number of days out of play were identified with nonparametric analyses (Spearman correlation, Mann-Whitney test, and Kruskal-Wallis test), because of the highly skewed distribution. For multivariate analyses, both multiple regression and logistic regression were used with stepwise selection of variables that were significant at α = 0.10 in the bivariate analysis. Factors associated with hospitalization were identified with χ2 tests for bivariate analysis and logistic regression for multivariate analysis. Only variables that were significant at α = 0.10 were used in the logistic regression.

Injury rates per 100 game-positions were calculated to provide information on the risk of injury among the position categories. The denominator for these rates reflects the number of standard position players multiplied by the number of team-games (3826 team-games) during the study period. For example, there is one quarterback position for each team in each game (game-position exposure = 1 × 3826). For the offensive line, there are five positions (one center, two guards, and two tackles) per team (game-position exposure = 5 × 3826). Injury rates per 1000 plays were also calculated for rushing, passing, kickoff, and punt plays, on the basis of the number of plays in the NFL for the study period (34). When type-of-play data were considered, proportions for the various types were based on regular-season games (16 games), because the NFL tracks the number of rushing, passing, kickoff, and punt plays. This allowed a comparison of the relative risks of injury for the four most common types of plays in the NFL.

RESULTS

Concussion in the NFL

During the 1996 to 2001 NFL seasons, there were 787 reported cases of MTBI in 3826 team-games (1913 games). This total included all preseason, regular-season, and playoff game-related concussions. Concussions occurred with an average incidence of 131.2 ± 26.8 concussions/yr and a rate of 0.41 concussion/game.

Table 2 presents the incidence of concussions according to player position in professional football. As a group, the offensive team experienced the highest frequency of concussions in the NFL. Individually, the position group most often associated with concussion was the defensive secondary (18.2%), followed by the kicking unit (16.6%) and the wide receivers (11.9%). When the injury rates per 100 game-positions were adjusted for the number of persons in each position group, the relative risk of concussion was highest for quarterbacks (1.62 concussions/100 game-positions), followed by wide receivers (1.23 concussions/100 game-positions), tight ends (0.94 concussion/100 game-positions), and defensive secondaries (0.93 concussion/100 game-positions). The positions at low risk were punter, return unit, kicker, and holder. When the quarterback, running backs, wide receivers, and defensive secondary were grouped as “backs” and the offensive and defensive linemen were grouped as “linemen,” the backs demonstrated nearly 3 times the relative risk of concussion.

TABLE 2.

Incidence of mild traumatic brain injury according to player position in National Football League games

Table 3 presents the types of plays in which concussions occurred. The highest frequency of injury was in passing plays (35.8%), followed by rushing plays (31.3%), kickoffs (15.9%), and punts (9.5%). Table 4 indicates that, when the injury rates per 1000 plays in regular-season games were considered, the relative risk of concussion in kickoff plays (9.29 concussions/1000 plays) was more than 4 times the risks in rushing and passing plays and 2.5 times the risk in punt plays (3.86 concussions/1000 plays). Kickoffs and punts were associated with significantly higher injury rates than were rushing or passing plays.

TABLE 3.

Types of play for concussed players in National Football League games

TABLE 4.

Injuries according to type of play for concussed players during regular-season National Football League games

Whereas Tables 3 and 4 summarize team activities at the time of the concussion, Figure 1 indicates the player activities during injury. The highest injury frequency was noted for players who were tackling (31.9%) or being tackled (28.6%). When data were combined, concussions were more often associated with tackles (60.5%) than with blocks (29.5%).

FIGURE 1.

Numbers of concussions associated with tackling, blocking, and other activities in NFL plays.

FIGURE 1.

Numbers of concussions associated with tackling, blocking, and other activities in NFL plays.

Figure 2 presents the percentages of concussions associated with different objects delivering the blow, from the epidemiological sample. The majority of concussions involved a strike by another player's helmet (67.7%). The remainder involved impact with the ground (11.4%) or impact by other body regions of a striking player (20.9%).

FIGURE 2.

Impacts from various objects associated with concussions in NFL games.

FIGURE 2.

Impacts from various objects associated with concussions in NFL games.

Table 5 summarizes the symptoms for concussed players in the 1996 to 2001 seasons. The median number of symptoms recorded for a concussion was 2 (range, 0–12). The three most common symptoms were headaches (55.0%), dizziness (41.8%), and blurred vision (16.3%). The least common symptoms were somatic complaints. There was considerable overlap between symptoms related to cognitive and memory problems, with 21.2% of the concussed players demonstrating both cognitive and memory problems. Overall, 45.9% of the concussed players experienced either cognitive or memory problems or both. No sensory or motor abnormalities were reported for any player, and nystagmus was the only cerebellar abnormality.

TABLE 5.

Initial symptoms for concussed players in National Football League gamesa

Table 6 summarizes the initial medical actions performed after injury. Only 16.1% of the players returned immediately to the game, including players who were evaluated on the field and returned to play and players who withdrew from the game for a few plays and then returned to play. The rest of the players rested for an extended period of time and returned in the same game (35.6%) or were removed from play and did not return to the game (44.0%). Nineteen players (2.4%) were removed from the game and hospitalized. Further analysis of data for the nonhospitalized players revealed strong associations between the presence and number of each symptom category and the action taken. Logistic regression was used to identify independent associations between the presence of categories, the number of symptoms, and the action taken. The model identified three variables as being independently associated with the action taken. The factors initially associated with resting of a player included the number of symptoms, any general symptom, any memory problem, any cognitive problem, and any somatic complaint. The multivariate logistic regression analysis identified the number of symptoms (odds ratio [OR] = 1.32 for each additional symptom), any memory problem (OR = 2.09), and any cognitive problem (OR = 3.41) as independent risk factors for resting of a nonhospitalized player.

TABLE 6.

Initial actions taken by team physicians for concussed players in National Football League games

Figure 3 presents the percentages of players returning to practice and games as a function of days lost after concussion in NFL games. Ninety-two percent of concussed players experienced ≤6 days lost and 97% experienced ≤9 days lost after MTBIs; 56.5% of concussed players experienced no days out of play. However, during the study period, 7% of concussed players experienced ≥7 days and 1.8% experienced >14 days out of play. Loss of consciousness occurred in 58 of 623 reported MTBI cases (9.3%) and involved more lost days, with 69% of players experiencing ≤6 days lost and 88% experiencing ≤9 days lost. Concussed players who sustained unconsciousness averaged 5.0 ± 7.5 lost days, which was 2.6 times longer than the time for players with MTBIs without loss of consciousness (1.9 ± 5.3 d).

FIGURE 3.

Percentages of players returning to practice and games after concussions with or without unconsciousness.

FIGURE 3.

Percentages of players returning to practice and games after concussions with or without unconsciousness.

Factors initially associated with increased days out of play included the number of symptoms, any general symptom, any memory problem, any cognitive problem, and any somatic complaint. The position played, type of play, action, or object of impact was not associated with days out of play. Although the results of the multivariate regression model suggested that only the number of symptoms was associated with an increased number of days out of play [F (1,785) = 46.27, P < 0.001], the analysis comparing any days missed with no days missed identified the number of symptoms (OR = 1.22 for each additional symptom), any general symptom (OR = 1.62), any somatic symptom (OR = 1.63), and any cognitive problem (OR = 2.35) as independent risk factors.

Comparison with the Video Analysis Cases

Figure 2 also indicates the percentages of concussions associated with different objects delivering the blow in the 182 video analysis cases reported by Pellman et al. (37,38). The comparison demonstrated almost identical distributions of objects delivering the blow, except for a lower incidence of impacts by the knee. There was greater involvement of quarterbacks, flankers, and split ends in the video cases, which is related to the greater chances of multiple video views of those players in game coverage. In addition, a greater proportion of video cases involved hospitalization.

DISCUSSION

This study is unique in many ways. Data were prospectively collected during a 6-year period, and a broad definition of MTBI was used. The pilot epidemiological work in the 1995 season helped refine the methods, and further efforts were taken to maximize the consistency and accuracy of MTBI reporting for this population. The study used a standardized reporting form to improve the team-to-team consistency of findings. The committee monitored the results reported with these standardized forms and adjusted the forms as needed to improve data collection. All subjects were evaluated by physicians and underwent follow-up physician evaluations until they returned to play. The forms were completed by physicians; therefore, there is increased confidence in the medical validity and reliability of the information collected. In addition, all patients underwent certified athletic trainer evaluations, and follow-up reports were included in the results. The report forms were completed contemporaneously with the examinations and were therefore not subject to the vagaries of recall at a later date.

As with any research project that includes epidemiological data from multiple recording sites, there are specific limitations to the data. Numerous individual team physicians evaluated the injured players and completed the data collection forms. Whenever there are multiple data recording sites and multiple data recorders, the question of interobserver reliability must be addressed. The documentation process was designed prospectively and included specific criteria for the inclusion of an MTBI, to minimize the effect of interobserver reliability. With these procedures in place, the overall effect of the diversity of the physicians and recording conditions for each team is not known. Despite efforts to ensure a standardized definition of MTBI and to increase the team physicians’ and athletic trainers’ knowledge of MTBI, it remains unclear whether the project identified all cases of MTBI during the study period. The well-known reluctance of professional athletes to report their injuries to medical personnel might have prevented the reporting of some MTBI cases.

Every player in the NFL was a potential subject for the study. Every player with a reported MTBI was evaluated and underwent follow-up evaluations if unable to return to play on the day of the injury. None of the players with MTBIs in the study was lost to follow-up monitoring. Another factor that makes this study unique is the fact that the number of players exposed to head injury during the 6-year period is known; therefore, relative risks could be determined according to player positions. There are no other studies with precise exposure data. The importance of the exposure data is that they enable the calculation of MTBI incidences in this population. The incidence data are not based on estimates or extrapolations, as in previous reports of incidences of MTBI, cerebral concussion, or serious head injury.

Because case management was not mandated by the committee, the medical course of the subjects reflects the true natural history of MTBIs among professional football players during this 6-year period. These epidemiological data and the report of the natural history of MTBIs among professional football players can serve as baseline findings for future monitoring. Because the monitoring of MTBIs will continue for the foreseeable future, future trends in MTBI incidence and severity in the NFL will be determined. This continuing effort will permit a study of the effects of newer protective equipment, rule changes, and perhaps even new medical treatments for head injuries on the incidence and severity of MTBIs. Such information should be of great value to professionals and students involved in brain injury. The applicability of these findings to athletes playing football at other levels (college, high school, or pee-wee), athletes participating in other contact sports, and/or the general population might be questioned. Certainly the players in this study are in a unique situation. They are highly trained, highly motivated, highly skilled athletes in excellent physical condition. They play with the best protective equipment available, which is maintained in the most efficient professional manner possible. They have the best training staff available to them on a daily basis. High-quality, personalized, medical care is also available. The players are monitored on a daily basis by athletic trainers and physicians who know them personally. However, despite these unique conditions, the players sustained MTBIs that seemed to be generally similar to the MTBIs experienced by other athletes and by the general population. This suggests that many of these findings are applicable to other athletes and the general population.

There have been previous attempts to study the epidemiological features of MTBI in the general population, among football players, and among other athletes. A brief review of those previous studies indicates the unique nature of this study. There have been many studies of head injury incidences in the general population (14). Those studies all yielded estimates of incidences, because the data were retrospectively collected from reviews of hospital records or insurance records; the values were then extrapolated to approximate incidences in the entire population. Such reported annual incidences of mild head injuries ranged from 131 cases/100,000 population to 511 cases/100,000 population. None of those studies relied on data collected from treating physicians, and none presented detailed physical findings.

There have been numerous attempts to epidemiologically study traumatic brain injury among football players. Gerberich et al. (16) reported on the incidence of football concussions in high schools in Minnesota. The data were collected with questionnaires sent to high school coaches and players. There were no physician reports. The data were all collected retrospectively. Torg et al. (43,44) reported on the National Football Head and Neck Injury Registry. The criteria for inclusion in the registry were not those of MTBI. The criteria designated serious injuries involving hospitalization, surgery, paralysis, or death. Data collection was clearly retrospective, from reports completed by high school principals, athletic trainers, and members of the American College of Sports Medicine at the end of each season. Information was also collected by a news-clipping service. The later years of the study reportedly used prospective data collection, but most of the data collection was still performed at the end of the season, with forms completed by the persons noted above, and the inclusion criteria still designated only serious brain injuries. Torg et al. (44) mentioned earlier work by Schneider (41a), who collected cases of intracranial hemorrhage, not MTBI, among high school and college football players between 1959 and 1963. There were no detailed reports from physicians. There were no reports of symptoms of MTBI, and obviously the criterion of intracranial hemorrhage is not indicative of MTBI.

Powell and Barber-Foss (39) reported on brain injuries among high school athletes, in a prospective study. The data were provided by athletic trainers who volunteered for the study. A standardized reporting form was used. Of the athletic trainers who volunteered, a fraction were selected to participate in the study, on the basis of representation of high schools of different sizes and geographical distribution throughout the United States. Reportable injuries included those that caused the cessation of customary participation in the current session of play. This criterion is somewhat broad and clearly includes many cases of MTBI. The reports were based on evaluations by athletic trainers, not physicians. An important strength of the study by Powell and Barber-Foss (39) was that incidents that occurred during practice and those that occurred during games were included. There was no report of clinical symptoms experienced by the players in the study. The total number of high school players in the United States was estimated from various sources; therefore, the incidence of traumatic brain injury could only be estimated from the data in the study.

Clarke (5) reported on the National Athletic Injury Report System, which was established in 1975. This report system was retrospective in nature. Concussion was defined as an incident of disorientation caused by trauma that required cessation of play. There were no reports of clinical symptoms. The reports were not based on physician examinations. There have been two reports of the incidence of concussions among Canadian football and soccer players (9,10). Both studies were based solely on retrospective survey questionnaires completed by the players themselves. There were no physician reports or examinations. Both studies reported very high incidences of concussion symptoms among the athletes.

There have been a few attempts to study brain injury in other sports. Gerberich et al. (15) reported an epidemiological study of high school hockey players in Minnesota during the 1982/1983 season. Only 12 high school teams participated. The study was retrospective and was not based on physician reports. There was no report of clinical symptoms observed among the injured players. Boxers have also been studied on a few occasions. Larsson et al. (25) examined 44 Swedish amateur boxers before and after matches. The boxers underwent neurological examinations, including brief bedside cognitive testing. That was a very small study that did not establish the incidence of concussion in the population of boxers. The report did not provide much detail regarding the clinical symptoms that were observed for the participating boxers.

McCown (31) retrospectively analyzed results from New York State Athletic Commission records on professional boxers who were examined between 1950 and 1958. The report indicated that 11,103 boxers were examined, 325 boxers were knocked out, 789 experienced technical knockouts, and 10 required hospitalization. The article included a few detailed case reports of injured boxers but did not include any systematic review of symptoms observed among the boxers who were evaluated. There was also no indication of the course of the patients’ symptoms after knockouts or technical knockouts.

Kaplan and Browder (22) evaluated 1043 professional boxers with electroencephalographic studies, ringside examinations, and evaluations of fight films. The authors did not specifically list the clinical findings, but they indicated that no neurological abnormalities were noted during ringside assessments or examinations in the training room after the fight, even among the fighters who lost in knockouts. No additional details regarding the clinical findings were provided.

Enzenauer et al. (12) reported on the incidence of boxing injuries in the United States Army between 1980 and 1985. That report was based on a retrospective review of Army hospital records. The only data available to the authors were the diagnoses, lengths of stay, and procedures performed. There were no clinical details regarding the symptoms. Because the data were collected from hospital records, only more serious head injuries were included and the report undoubtedly did not detect most cases of MTBI among Army boxers. Reviews of those earlier studies support the conclusion that our study is unique.

The results of this study help validate the information presented in the earlier study of the biomechanical features of concussions determined with video analyses and laboratory reconstructions of selected NFL MTBIs and severe head impacts (37,38). One concern regarding the selection of the cases for video analysis was that the cases might not have been representative of most, if not all, concussions observed among NFL players. The analytical method required multiple camera angles and clear yard line and sideline markers for accurate determination of impact conditions. Therefore, the cases analyzed with the video techniques tended to be those that occurred in the open field. Such cases tended to involve impacts between wide receivers and defensive backs, impacts between wide receivers and the ground, players striking the quarterback, and special-team players running in the open field and striking other players in kicking plays. This study indicates that these are the same players who have the highest risk of sustaining MTBIs among all NFL players. The data indicate that quarterbacks, wide receivers, defensive backs, and special-team players on kicking units are more likely to sustain MTBIs than are players such as offensive and defensive linemen. This indicates that the cases selected for video analysis in the earlier studies were representative of many, if not all, NFL concussive events. Furthermore, comparison of the symptoms indicated that the video sample was reasonably representative.

Some of the results in this report can be understood in the context of the biomechanical data determined in the earlier studies (37,38). The data indicate that quarterbacks have the highest risk of MTBI among all NFL players. Quarterbacks are representative of immobile or slowly moving players who are struck at high velocity by other players, often in situations in which the quarterback is unaware of the approaching player. The biomechanical data indicate that, in such situations, the high velocity of the striking player is transferred to the struck player (the quarterback), causing large changes in velocity and acceleration of the quarterback's head.

Defensive backs and wide receivers are the next most likely groups of NFL players to sustain clinical MTBIs. These players are moving at high speeds. They are often struck by more than one teammate or opposing player in high-velocity, high-acceleration impacts during tackling or blocking. Wide receivers are often struck in midair and fall backward, landing with the back of the helmet against the ground. As indicated by the biomechanical data, these high-velocity impacts often produce head accelerations exceeding the tolerance levels and thus frequently result in clinical MTBIs. Offensive and defensive linemen have relatively lower risks of MTBI than do the aforementioned players. Offensive and defensive linemen tend to move at slower velocities for shorter distances; therefore, when they are struck in the head, the velocities and accelerations are lower than those observed among quarterbacks, wide receivers, and defensive backs. Because of these lower-acceleration impacts, lineman are less likely to experience impacts exceeding the tolerance levels and are less likely to sustain clinical MTBIs, except when they are running at full speed during kickoffs and punts.

One of the results might seem inconsistent with the biomechanical data. These data indicate that tackling players are somewhat more likely to sustain MTBIs than are players being tackled. Our biomechanical data indicated that the striking player does not sustain MTBI but the player being struck sustains MTBI because of the velocity being transferred from the striking player to the struck player. It might initially be presumed that the tackling player is always the striking player, but closer analysis indicates that this is often not the case. The tackling player, especially a relatively smaller defensive back, is often struck by a quickly moving, charging, offensive ball carrier, who becomes the striking player although he is being tackled. Therefore, the higher accelerations and velocities are often transferred to the tackling player. Furthermore, the tackling player is often struck by other tackling players from his own team or by blocking players from the opposing team during the act of tackling.

The tackler is usually focused on the ball carrier and often does not see the blocking opponent or his teammate who is also attempting to make a tackle. In these collisions, the tackling player often experiences a high-velocity impact, sustaining high accelerations to the head that result in concussion. Therefore, even this apparent inconsistency can be understood in terms of the biomechanical data presented by Pellman et al. (37,38).

As indicated in Table 5, the most common initial symptoms for concussed players were headaches, dizziness, memory problems, cognitive problems, and somatic complaints. Headaches were observed for 55.0% of concussed NFL players. These results are consistent with the results observed in previous studies. Headaches occur among 30 to 90% of patients after concussions (14,33). In recent studies of MTBIs among athletes, posttraumatic headaches were noted in 40 to 86% of cases (7,18,32). Our results were within that range and are consistent with those findings.

Dizziness, including vertigo, was observed for 45.7% of the NFL players with concussions. Dizziness has been frequently reported in studies of nonathletes with closed-head injuries. One study demonstrated that 53% of nonathletes complained of dizziness after mild head injuries (14,27). Dizziness and vertigo are usually indicative of vestibular system dysfunction. This dysfunction is often thought to be of peripheral origin, but it is not clear whether some of the dizziness and vertigo might be of central origin. Earlier reports suggested that mild head injuries could cause benign positional vertigo. Nystagmus was infrequently noted among the NFL players. On the basis of examinations of boxers in the ring immediately after knockouts, nystagmus might be more frequently observed among concussed NFL players. The low incidence of nystagmus in this study could be related to the fact that very few of the NFL players experienced loss of consciousness or the fact that the players were not examined immediately after impact.

Blurred vision was observed for 16.3% of the study population. A previous study demonstrated a 14% incidence of blurred vision among nonathletes with concussions (33). Double vision was observed for 2% of the concussed NFL players. No additional details regarding the double vision were recorded in the survey, but double vision attributable to IVth cranial nerve palsy is well known to occur after mild head injury in the nonathlete population (24). Photophobia (sensitivity to light) was observed for 4.1% of the NFL players with concussions. In an earlier study, photophobia was observed for 7.2% of nonathletes examined 14 days after mild head injuries (14,17). Photophobia may be part of a posttraumatic migraine syndrome.

Memory and cognitive problems were observed for a total of 45.9% of the NFL players after concussion. Memory problems were observed for 39.5% of the players and cognitive problems for 27.6%. Other sources indicated that retrograde amnesia is one of the most common forms of memory dysfunction with MTBI (27,28,30). Diagnostic approaches for assessment of memory function and cognitive problems were not uniform among the team physicians, which represents a potential error in the accuracy of reporting of such signs. Of the overall sample, 25.5% of players demonstrated difficulties with immediate recall in the initial examination, 8.0% were disoriented with respect to time, 5.1% were disoriented with respect to place, and 2.9% were disoriented with respect to person. Impairment of immediate recall was thus much more frequent than disorientation among the NFL players. These results confirm our clinical experience that it is not enough to ask a concussed athlete what year, month, or day it is to ascertain whether he is experiencing cognitive difficulties. It is necessary to specifically test for immediate recall in the initial patient examination after MTBI, to accurately assess the nature and extent of the injury. This is an important message for athletic trainers, sports medicine physicians, and others who examine athletes on the sidelines after mild head injuries. Team physicians were not given standardized questions or formats for assessment of information-processing ability or memory.

The 39.5% incidence of memory problems was approximately equally divided between retrograde amnesia, information-processing problems, attention problems, and anterograde amnesia. These incidences are consistent with the findings of previous studies of nonathletes and athletes after concussions. In one study of nonathletes examined 4 weeks after mild head injuries, 19% complained of memory losses and 21% complained of concentration difficulties (14,33). There is extensive literature on the neuropsychological sequelae of MTBIs among high school and college football players, as well as other athletes (6,8,11,13,14,19,46). Those articles reported significant abnormalities on cognitive and memory tests among concussed athletes. Our results are consistent with those earlier reports.

Somatic complaints were observed for 20.1% of concussed NFL players. The most common somatic complaints were fatigue, anxiety, personality changes, irritability, and sleep disturbances. These results are consistent with previous studies of symptoms noted after mild head injuries among nonathletes. It has been reported that 50 to 84% of patients experience these types of symptoms after cerebral concussion (14, 4042). Fatigue was noted for 29% of patients evaluated 4 weeks after mild head injuries (14,33). Disrupted sleep patterns have also been reported (14,35). We are not aware of studies of the incidences of somatic complaints among other groups of athletes after mild head injuries. However, the MTBI results are consistent with findings reported for the nonathlete population, indicating the importance of inquiries about somatic complaints during evaluations of athletes after concussions. Complaints of fatigue, anxiety, personality changes, irritability, and sleep disturbances can be significantly disabling to the patient and must be taken seriously by the evaluating physician.

There have been a few studies of some symptoms after acute head injuries among amateur and professional boxers. Kaplan and Browder (22) reported normal neurological examination results for boxers after knockouts but noted that the “loss of awareness of one's surroundings” was often observed for fighters after concussions. Blonstein and Clarke (2) evaluated 29 amateur boxers who had been knocked out during the 1955/1956 boxing season in England. They did not report specific values, but they indicated that amnesia was frequently observed. Kaplan and Browder (22) and Blonstein and Clarke (2) also performed electroencephalographic examinations of fighters after cerebral concussion, but the comments regarding those findings did not mention clinical symptoms. In another report addressing head injuries in boxing, Winterstein (47) reported that nearly one-half of the boxers he examined described instances of anterograde amnesia. No additional details regarding the number of boxers examined or the nature or extent of the evaluations were provided in that report. It seems clear that the symptoms observed among concussed NFL players were generally consistent with the symptoms previously noted for athletes and nonathletes after MTBIs. These findings for a group of professional athletes are certainly more detailed and comprehensive than those reported in prior studies.

Only one NFL player experienced a seizure after MTBI. The seizure occurred in the locker room approximately 30 minutes after the injury and could therefore be classified as an early posttraumatic seizure. The patient made a full recovery. Posttraumatic seizures represent a rare occurrence in the general population after MTBI. They occur for 0.8 to 2.3% of patients during the first 1 week after mild head injuries (1,14,26).

For the great majority of concussed NFL players, MTBIs did not cause prolonged disability or prolonged absence from play. Figure 3 indicates that 56.5% of the players (445 players) returned to play on the day of the injury and 97.1% (764 players) returned to play by day 9 after the injury. Only 23 players (2.9%) missed more than 9 days before returning to play. This indicates that most MTBIs in the NFL are self-limiting and players recover fully and spontaneously in a short time. Table 6 indicates that more than one-half of the concussed players returned to play during the same game, either immediately (16.1%) or after a period of rest (35.6%). Conversely, 44.0% did not return to the same game. The data did not reveal some clinically pertinent information, such as the number of players who returned to play with symptoms, but they allowed an analysis of the types of symptoms that were more likely to be associated with removal from play. The multivariate logistic regression analysis of nonhospitalized players identified the number of symptoms (OR = 1.32 for each additional symptom), any memory problem (OR = 2.09), and any cognitive problem (OR = 3.41) as independent risk factors for resting of a player. The MTBI Committee did not intend to interfere with clinical decision-making by the individual medical staffs, and it must be assumed that there were variations in the treatment of the injuries. Because a significant percentage of players returned to play in the same game and the overwhelming majority of players with concussions were kept out of football-related activities for less than 1 week, it can be concluded that MTBIs in professional football represent mild injuries, in the context of the wide spectrum of diffuse brain injuries. This suggests that MTBIs in the NFL represent injuries in which symptoms resolve within a short time in the vast majority of cases.

Only 9.3% of the NFL players in this study experienced loss of consciousness as a result of severe concussive head impacts. This small number is consistent with clinical experience indicating that loss of consciousness is not a common occurrence with football-related MTBIs. It is important for all physicians who care for athletes with head injuries to know that most concussions they treat are not associated with loss of consciousness. Our data also demonstrated that concussed football players who sustained a loss of consciousness returned to play significantly later than did those who did not sustain a loss of consciousness. There are a few possible explanations for this finding. It could be that players who experience a loss of consciousness exhibit more severe symptoms or different patterns of symptoms than do those who do not lose consciousness. Players who lose consciousness may exhibit more abnormalities in examinations than do their counterparts who do not lose consciousness. It also may take longer for the symptoms of concussion to resolve among players who experience a loss of consciousness, compared with players who do not sustain a loss of consciousness.

It is also possible that the symptom of loss of consciousness itself is enough to make the treating physician more concerned regarding the potential seriousness of the injury and thus delay return to play longer, although the symptoms are not different from those for players who did not sustain a loss of consciousness. The available data do not allow differentiation between these possibilities. However, a more detailed analysis of the subgroup of athletes who sustained a loss of consciousness is planned for a subsequent report.

MTBIs are relatively common injuries sustained by professional football players in game situations. As expected, the risk of injury is greatest for players whose positions more frequently involve high-velocity impacts (wide receivers and defensive backs). The data demonstrated that the quarterback is the most at-risk position, after adjustment for the number of positions on the field (Table 2). Rule changes have been implemented in the past to help protect this position from the potentially increased risk (e.g., pre- and post-throwing positions and blind-side impacts).

These data do not differentiate between initial and recurrent concussions. Repeated concussions are of concern because of the potential risks of chronic deficits and prolonged postconcussion syndrome. NFL players have occasionally been forced to retire because of prolonged postconcussion syndrome. However, most players return to play very soon after concussion and long-term sequelae after MTBI are quite rare in the NFL.

The results of a unique, prospective, 6-year study of the epidemiological features, natural history, and clinical features of MTBIs in professional football were presented. The results were analyzed with respect to player position and type of football activity (i.e., blocking, being blocked, tackling, or being tackled). Details of the clinical findings for MTBI were reported for 787 cases, and an attempt was made to correlate those findings with the biomechanical data gleaned from video analyses and impact reconstructions of concussions, as previously presented by Pellman et al. (37,38). These findings should assist physicians in the diagnosis, treatment, and counseling of patients who sustain MTBIs.

Acknowledgments

The authors are members of the Mild Traumatic Brain Injury Committee, National Football League, New York, New York. The efforts of another committee member, Ronnie Barnes, A.T.C., are appreciated. The NFL MTBI Committee is chaired by Dr. Elliot Pellman and includes representatives from the NFL Team Physicians Society and the NFL Athletic Trainers Society, NFL equipment managers, and scientific experts in the areas of traumatic brain injury, biomechanics, basic science research, and epidemiology. None of the committee members has a financial or business relationship that poses a conflict of interest with respect to the research on concussion in professional football. The MTBI Committee gratefully acknowledges the insight of Commissioner Paul Tagliabue in forming the committee and issuing a charge to scientifically investigate concussion and the means to reduce injury risks in football. The efforts of Dorothy Mitchell, former counsel for the NFL, are also gratefully acknowledged. She worked tirelessly to initiate the MTBI research. Although she left the NFL in the middle of the program, her efforts paved the way for successful completion of the research. The encouragement and support of Jeff Pash and Peter Hadhazy of the NFL are appreciated. The committee also appreciates the contributions of all of the NFL team physicians and athletic trainers who completed the MTBI report forms and the players who consented to participate in the epidemiological study through blinded identification in the MTBI database. We thank the staff members at Med Sports Systems for their efforts in managing the data flow among the various aspects of the project. Without their support, the project could not have been completed. Finally, the committee appreciates the assistance of Prof. Cynthia Arfken, Wayne State University, in the statistical/epidemiological analysis of the concussion database. Funding for this research was provided by the NFL and the NFL Charities. The NFL Charities are funded by the NFL Player's Association and League. Their support and encouragement for research on concussion are greatly appreciated.

REFERENCES

1.
Annegers
J
,
Grabow
J
,
Groover
RV
,
Laws
ER
Jr
,
Elveback
LR
,
Kurland
LT
Seizures after head trauma: A population study
.
Neurology
 
30
:
683
689
,
1980
.
2.
Blonstein
J
,
Clarke
E
Further observations on the medical aspects of amateur boxing
.
Br Med J
 
16
:
362
364
,
1957
.
3.
Cantu
RC
Classification and clinical management of concussion
, in
Bailes
JE
,
Day
AL
(eds):
Neurological Sports Medicine: A Guide for Physicians and Athletic Trainers
 .
Park Ridge
,
AANS
,
2001
, pp
25
33
.
4.
Centers for Disease Control and Prevention
:
Sport-related recurrent brain injuries
.
MMWR Morb Mortal Wkly Rep
 
46
:
224
227
,
1997
.
5.
Clarke
KS
Epidemiology of neurologic injuries in sports
, in
Jordan
B
,
Tsairis
P
,
Warren
RF
(eds):
Sports Neurology
 .
Aspen
,
Aspen Publishers
,
1989
, pp
63
74
.
6.
Collie
A
,
Darby
D
,
Maruff
P
Computerized cognitive assessment of athletes with sports related head injuries
.
Br J Sports Med
 
35
:
297
302
,
2001
.
7.
Collins
MW
,
Field
M
,
Lovell
MR
,
Iverson
GL
,
Johnston
KM
,
Maroon
J
,
Fu
FH
Relationship between post-concussion headache and neuropsychological test performance in high school athletes
.
Am J Sports Med
 
31
:
168
173
,
2003
.
8.
Collins
M
,
Grindel
S
,
Lovell
M
,
Dede
D
,
Moses
D
,
Phalin
B
,
Nagle
S
,
Wasik
M
,
Cordry
D
,
Dougherty
KM
,
Sears
S
,
Nicolette
G
,
Indelicato
P
,
McKeog
D
Relationship between concussion and neuropsychological performance in college football players
.
JAMA
 
282
:
964
970
,
1999
.
8A.
Congress of Neurological Surgeons
:
Proceedings of the Congress of Neurological Surgeons in 1964: Report of the Ad Hoc Committee to Study Head Injury Nomenclature
.
Clin Neurosurg
 
12
:
386
394
,
1966
.
9.
Delaney
JS
,
Lacroix
VJ
,
Leclerc
S
,
Johnston
KM
Concussions during the 1997 Canadian Football League season
.
Clin J Sports Med
 
10
:
9
14
,
2000
.
10.
Delaney
JS
,
Lacroix
VJ
,
Leclerc
S
,
Johnston
KM
Concussions among university football and soccer players
.
Clin J Sports Med
 
12
:
331
338
,
2002
.
11.
Echemendia
R
,
Putukian
M
,
Mackin
S
,
Julian
L
,
Shoss
N
Neuropsychological test performance prior to and following sports related mild traumatic brain injury
.
Clin J Sports Med
 
11
:
23
31
,
2001
.
12.
Enzenauer
R
,
Montrey
J
,
Enzenauer
R
,
Mauldin
M
Boxing-related injuries in the US Army: 1980 through 1985
.
JAMA
 
261
:
1463
1466
,
1989
.
13.
Erlanger
D
,
Saliba
E
,
Barth
J
,
Almquist
J
,
Webright
W
,
Freeman
J
Monitoring resolution of postconcussion symptoms in athletes: Preliminary results of a web-based neuropsychological test protocol
.
J Athl Train
 
36
:
280
287
,
2001
.
14.
Evans
R
The post concussion syndrome and the sequelae of mild head injury
, in
Evans
R
(ed):
Neurology and Trauma
 .
Philadelphia
,
W.B. Saunders Co.
,
1996
, pp
91
116
.
15.
Gerberich
SG
,
Finke
R
,
Madden
M
,
Priest
JD
,
Aamoth
G
,
Murray
K
An epidemiologic study of high school ice hockey injuries
.
Childs Nerv Syst
 
3
:
59
64
,
1987
.
16.
Gerberich
S
,
Priest
J
,
Boen
JR
,
Staub
C
,
Maxwell
RE
Concussion incidence and severity in secondary school varsity football players
.
Am J Public Health
 
73
:
1370
1375
,
1983
.
17.
Gronwall
D
,
Wrightson
P
Delayed recovery of intellectual function after minor head injury
.
Lancet
 
2
:
605
609
,
1974
.
18.
Guskiewicz
K
,
Weaver
N
,
Padua
D
,
Garrett
W
Epidemiology of concussion in collegiate and high school football players
.
Am J Sports Med
 
28
:
643
650
,
2000
.
19.
Iverson
G
,
Lovell
M
Validity of impact for measuring the effects of sports related concussion
 .
Presented at the National Academy of Neuropsychology Annual Conference
,
Miami
,
February 25–March 3, 2002
.
20.
Johnston
KM
,
McCrory
P
,
Mohtadi
NG
,
Meeuwisse
W
Evidence-based review of sport-related concussion: Clinical science
.
Clin J Sports Med
 
11
:
150
159
,
2001
.
21.
Jordan
B
Head injury in sports
, in
Jordan
B
,
Tsairis
P
,
Warren
RF
(eds):
Sports Neurology
 .
Rockville
,
Aspen Publishing
,
1989
, pp
75
83
.
22.
Kaplan
H
,
Browder
J
Observations on the clinical and brain wave patterns of professional boxers
.
Jour AMA
 
156
:
1138
1144
,
1954
.
23.
Krauss
J
,
McArthur
D
Epidemiology of brain injury
, in
Evans
R
(ed):
Neurology and Trauma
 .
Philadelphia
,
W.B. Saunders Co.
,
1996
, pp
3
17
.
24.
Kwartz
J
,
Leatherbarrow
B
,
Davis
H
Diplopia following head injury
.
Injury
 
21
:
351
352
,
1990
.
25.
Larsson
LE
,
Melin
KA
,
Nordstrom-Ohrberg
G
,
Silfverskiold
BP
,
Ohrberg
K
Acute head injuries in boxers: Clinical and electroencephalographic studies
.
Acta Psychiatr Neurol Scand Suppl
 
95
:
2
42
,
1954
.
26.
Lee
S
,
Lui
T
Early seizures after mild closed head injury
.
J Neurosurg
 
76
:
435
439
,
1992
.
27.
Levin
H
,
Mattis
S
Neurobehavioral outcome following minor head injury: A three center study
.
J Neurosurg
 
66
:
234
243
,
1987
.
28.
Lovell
MR
,
Collins
MW
,
Iverson
GL
,
Field
M
,
Maroon
JC
,
Cantu
R
,
Podell
K
,
Powell
JW
,
Belza
M
,
Fu
FH
Recovery from mild concussion in high school athletes
.
J Neurosurg
 
98
:
296
301
,
2003
.
29.
Deleted in proof.
30.
Maroon
J
,
Lovell
M
,
Norwig
J
,
Podell
K
,
Powell
J
,
Hartl
R
Cerebral concussion in athletes: Evaluation and neuropsychological testing
.
Neurosurgery
 
47
:
659
672
,
2000
.
31.
McCown
I
Boxing injuries
.
Am J Surg
 
98
:
509
516
,
1959
.
32.
McCrory
PR
,
Ariens
M
,
Berkovic
SF
The nature and duration of acute concussive symptoms in Australian football
.
Clin J Sports Med
 
10
:
235
238
,
2000
.
33.
Minderhoud
J
,
Boelens
M
,
Huizenga
J
,
Saan
RJ
Treatment of minor head injuries
.
Clin Neurol Neurosurg
 
82
:
127
140
,
1980
.
34.
National Football League
:
Official NFL Record and Fact Book
 .
New York
,
NFL Publications
,
1997–2002
.
35.
Parsons
L
,
Verbeek
D
Sleep-awake patterns following cerebral concussion
.
Nurs Res
 
31
:
260
264
,
1982
.
36.
Pellman
EJ
Background on the National Football League's research on concussion in professional football
.
Neurosurgery
 
53
:
797
798
,
2003
.
37.
Pellman
EJ
,
Viano
DC
,
Tucker
AM
,
Casson
IR
Concussion in professional football: Location and direction of helmet impacts–Part 2
.
Neurosurgery
 
53
:
1328
1341
,
2003
.
38.
Pellman
EJ
,
Viano
DC
,
Tucker
AM
,
Casson
IR
,
Waeckerle
JF
Concussion in professional football: Reconstruction of game impacts and injuries
.
Neurosurgery
 
53
:
799
814
,
2003
.
39.
Powell
J
,
Barber-Foss
K
Traumatic brain injury in high school athletes
.
JAMA
 
282
:
958
963
,
1999
.
40.
Rimel
R
,
Giordani
R
,
Barth
JT
,
Bool
TJ
,
Jane
JA
Disability caused by minor head injury
.
J Neurosurg
 
9
:
221
228
,
1981
.
41.
Rutherford
W
,
Merrett
J
,
McDonald
JR
Sequelae of concussion caused by minor head injuries
.
Lancet
 
1
:
1
4
,
1977
.
41A.
Schneider
RC
Head and Neck Injuries in Football: Mechanisms, Treatment and Prevention
 .
Baltimore
,
Williams & Wilkins
,
1973
.
42.
Slagle
D
Psychiatric disorders following closed head injury
.
Int J Psychiatry Med
 
20
:
1
35
,
1990
.
43.
Torg
R
,
Vegsa
J
The National Football Head and Neck Injury Registry: 14 year report on cervical quadriplegia, 1971 through 1984
.
JAMA
 
254
:
3439
3443
,
1985
.
44.
Torg
R
,
Truex
R
,
Quedenfeld
T
,
Burstein
A
,
Spealman
A
,
Nichols
C
The National Football Head and Neck Injury Registry: Report and conclusions 1978
.
JAMA
 
241
:
1477
1479
,
1979
.
45.
Tuohima
P
Vestibular disturbances after acute mild head injury
.
Acta Otolaryngol (Stockh)
 
359
[Suppl]:
1
59
,
1978
.
46.
Warden
D
,
Bleiberg
J
,
Cameron
K
,
Sun
W
,
Sparling
M
,
Cernich
A
,
Peck
K
,
Reeves
D
,
Walter
J
,
Uhorchok
J
,
Eckland
J
The effect of concussion history on cognitive performance following acute concussion
.
Neurology
 
60
[Suppl
1
]:
A362
,
2003
(abstr).
47.
Winterstein
C
Head injuries attributable to boxing
.
Lancet
 
2
:
719
720
,
1937
.

COMMENTS

In this article by Pellman et al., the third in a series, the authors detail the circumstances, causes, symptoms, and outcomes of game-related concussions in National Football League (NFL) players during a 6-year period. The authors found that quarterbacks, wide receivers, tight ends, and defensive secondary positions had the highest relative risk for concussion, and kickoffs and punt returns were the types of play that carried the highest relative risk for concussion. Importantly, they also showed that only 9% of concussed players lost consciousness, emphasizing the fact that in the great majority of cases, concussion does not equate with unconsciousness. Instead, the most common signs and symptoms were headache, dizziness, memory problems, and cognitive complaints. The authors have provided a comprehensive and in-depth analysis of this complex clinical problem. This work will allow focus to be placed on the most vulnerable players in the game to reduce the incidence and severity of concussion in football players at all levels. Further study is warranted on the long-term neurobehavioral effects of repeated concussions relative to concussion frequency, injury severity, and time interval between concussions.

Daniel F. Kelly

Los Angeles, California

This article is the third in a series of research observations supported by the NFL and performed through the NFL Committee on Mild Traumatic Brain Injury (MTBI) in response to safety concerns regarding head injuries in the NFL. The present observations represent the institutionalization of change in the former understanding of a concussion as an episode associated with loss of consciousness to the currently recognized definition as “an altered mental state regardless of duration and/or altered memory, regardless of duration or content that results from trauma.” The authors’ description and categorization of symptoms and signs commonly associated with concussion clearly establish these as the hallmarks in defining MTBI in sports.

During the 6-year study period, 3826 team games were analyzed for the incidence and severity of MTBI by player position. From preseason, regular, and playoff game-related concussions, there were 787 reported cases of MTBI. The incidence was only 0.41 concussion per NFL game. This seemingly low incidence perhaps highlights one of the few limitations of the study: the known reluctance of some athletes to report symptoms and signs. Also, with the numerous individual team physicians and trainers filling out the collection forms, interobserver reliability and lack thereof remains a question. In these incidents, 16.1% of players returned immediately to the game, 35.6% returned later in the game, and 44% were removed from further contact that day; 2.4% were subsequently hospitalized.

As one would expect, the risk of injury is greatest for players involved in high-velocity impacts, such as quarterbacks, wide receivers, and defensive backs. Because of the high vulnerability of quarterbacks, rule changes were implemented to help protect against blind-side impacts and prethrowing and postthrowing positions.

The great value in this study is the promulgation of a standard definition, the delineation of players at greatest risk, and the detailed analysis of symptoms and signs associated with MTBI. It is a prospective study that used a standardized reporting form in which all subjects were evaluated by physicians, and there was 100% participation from all players in the NFL, with none lost to follow-up. The NFL and the members of the MTBI Committee of the NFL are to be commended for their concise and succinct summary of 5 years of work.

Joseph C. Maroon

Pittsburgh, Pennsylvania

In their latest article investigating concussion during professional football games in the United States, Pellman et al. change their focus from biomechanics to an analysis of the positions most likely to sustain concussion and the most common signs and symptoms. The fact that quarterbacks were found to have the highest relative risk of concussion should come as no surprise to anyone who reads the sports pages between August and January. However, the finding that is perhaps most important for all practitioners (even those who are not sports fans) is that fewer than 10% of all concussions were complicated by loss of consciousness. It cannot be stated often enough to the general public and even to other medical practitioners that concussion can occur without loss of consciousness and, in fact, does so quite frequently.

Fortunately, in the vast majority of cases, MTBI in football players seems to resolve quickly. Most players are able to return to competition after a relatively brief period. The subset of patients who sustain loss of consciousness may experience more significant sequelae, and the authors report that they plan to present a more detailed analysis of these athletes in a future article. I look forward to reading that article. The authors also state that they plan to investigate the effects of recurrent concussions in players. Concerns about the effects of repeated concussions seem to have played a role in the retirement of some of the most prominent football players of recent years. The high visibility and high stakes of such decisions guarantee that the publication of that article will also be eagerly awaited.

Alex B. Valadka

Houston, Texas

This article represents the third part of an informative series regarding concussive injury in the NFL. A significant contribution is the documentation that players in the NFL with MTBI tend to return to play quickly without increasing the chance for subsequent sequelae. This counters an existing literature that tends to suggest that even minimally symptomatic players with MTBI have serious injuries. It is unfortunate that, to date, only a few articles discussing MTBI in general have been prospective in nature, especially given the burden of a 3.9 to 7.7% rate of MTBI in high school and college athletes each year in all sports. We agree that the use of standardized criteria for recording variables related to the injury is a significant contribution, especially as determined by a small, select group of trainers and/or physicians. The realization that 182 severe game impacts were the result of spearing should support the need for a more aggressive stance on the part of the NFL with regard to penalization of the offending player. The authors will be pursuing repeat concussion in a future article, given the significance of such injuries with regard to long-term and even career-ending sequelae.

As with any study of this sort, a number of issues remain problematic. Our experience has been that motivated players will universally underrepresent their injuries, given the will to return to play. Given the caliber and motivation of elite athletes, there is little chance that this will change. Thus, criteria to assess players with potential MTBI need to account for this variable. Close relationships between the trainers, players, and physicians are necessary to further reduce the impact of a player minimizing his injuries. In addition, close adherence to strict criteria of evaluation will further diminish such issues with regard to return-to-play considerations.

Our greatest concern is that the physicians and trainers truly represent the well-being and long-term health of the players, as opposed to the specific ball club or the NFL. It is unusual that neurosurgical coverage tends to be limited in the NFL, especially given the significant knowledge base regarding head injury that we harbor. In a recent Monday Night Football game, Jeff Garcia (quarterback of the San Francisco 49ers) probably sustained a concussion as the result of a open-field tackle. Replay analysis represented the tackle itself and Garcia's hesitancy to leave the field. Without the benefit of performing an examination of the player, any consideration of the degree of his injury is only conjecture. Conversely, the visual evidence was enough to suggest that he did receive a significant impact that was associated with a concussive injury. In any case, the rapidity with which he returned to play was of concern. It has been only a few years since Troy Aikman of the Dallas Cowboys was rendered unconscious on the playing field, only to return to play later in the game.

Burak Ozgur

Michael L. Levy

San Diego, California

The NFL's Committee on MTBI has analyzed clinical data regarding what they believe is a comprehensive capture mechanism of virtually all instances of concussion in the league during the 1996 to 2001 football seasons. This study has yielded information from 787 MTBIs that occurred throughout the playing seasons in 1913 games and consisted of both initial and follow-up evaluations. The findings included the fact that although the positions most often associated with concussion are the defensive secondary, followed by the kicking team members and wide receivers, the quarterback position has the highest relative risk for sustaining concussion. This confirms our previous conviction that those involved in high-speed collisions, that is, the quarterbacks, running backs, wide receivers, and defensive secondary players, are at greatest risk for sustaining concussion. However, depending on size, velocity, and other factors, the tackler may not always be the one imparting the greater force and may himself sustain the concussion. In addition, the quarterback is often the victim of a “blind-side” hit, being unaware of the closing angle, speed, and presence of the tackler, and is often a stationary target absorbing and not delivering a force vector to the opponent.

The findings also showed that most concussed football players had two symptoms, with headaches, dizziness, and blurred vision being the most common. Either cognitive or memory problems or both were seen in 46%, whereas, interestingly, there was no documentation of motor or sensory abnormalities in any players. It is also of interest to note that only a minority (16.1%) of MTBI players returned to play immediately within the same game, whereas 35.6% rested for an extended period before returning to the same game, and 44% were removed for the remainder of that contest. Loss of consciousness was present in 9.3%, and only 19 players (2.4%) required hospitalization. The return to play was rapid, because 92% of concussed players had 6 or fewer days away from play, and 97% were absent for fewer than 10 days.

This report contributes much to our understanding of the incidence and characteristics of football-related MTBI, at least at the highest level of competition. The methodology was inclusive, allowing exposure and incidence to be determined accurately, as well as presenting symptomatology. Although there are some limitations to their data, as the authors admit, this study involved examinations by physicians and good follow-up information, including the athletes’ ability to return to competition. It did not differentiate between initial and multiple concussions. The large number of concussions in highly skilled and conditioned athletes, the data related to playing position, physical examination and presentation, and completeness make this study unique. The findings also underscore the importance of assessing memory by testing for immediate recall in the sideline evaluation. The authors have presented a thorough study that documents many important features of athletic MTBI at the professional level.

Julian E. Bailes

Morgantown, West Virginia

Representation of the constellation Pegasus, visible in both hemispheres, right ascension: 22 hours, declination: +20°. Bellerophon, a hero in Greek mythology, was the second human to tame and ride Pegasus. Bellerophon was given several difficult tasks; his success in accomplishing those tasks made him proud, and in his pride he attempted to fly Pegasus to Olympus, as if he were an immortal. His presumption angered Zeus, who sent a horsefly to bite Pegasus. Pegasus reared, and Bellerophon fell to earth; thereafter, he spent the rest of his life as a lame, blind wanderer. Pegasus was accepted at Olympus, where he carried Zeus’ thunderbolts.

Representation of the constellation Pegasus, visible in both hemispheres, right ascension: 22 hours, declination: +20°. Bellerophon, a hero in Greek mythology, was the second human to tame and ride Pegasus. Bellerophon was given several difficult tasks; his success in accomplishing those tasks made him proud, and in his pride he attempted to fly Pegasus to Olympus, as if he were an immortal. His presumption angered Zeus, who sent a horsefly to bite Pegasus. Pegasus reared, and Bellerophon fell to earth; thereafter, he spent the rest of his life as a lame, blind wanderer. Pegasus was accepted at Olympus, where he carried Zeus’ thunderbolts.